KRAS-retroviral fusion transcripts and gene amplification in arsenic-transformed, human prostate CAsE-PE cancer cells

CAsE-PE cells are an arsenic-transformed, human prostate epithelial line containing oncogenic mutations in KRAS. In this in vitro transformation model, arsenic exposure produced a highly overexpressed, mutated KRAS allele in CAsE-PE cells compared to the immortalized parent cells, RWPE-1, expressing normal KRAS transcript. We previously observed an increased KRAS copy number in CAsE-PE, supporting a process of gene amplification of unclear origin. Here, we endeavored to determine the flanking genomic and transcriptomic sequences to KRAS in CAsE-PE cells for insight into its gene amplification. A comparison of DNA-Seq and RNA-Seq reads showed coincident reads aligned to all KRAS exons in CAsE-PE cells, while only RNA-Seq reads were observed in RWPE-1 cells. We searched for KRAS fusions in DNA and RNA sequencing data finding a portion of reads aligning to KRAS and viral sequence. After generation of cDNA from total RNA, we designed short and long probes to hybridize to KRAS cDNA and then sequenced released fragments by PacBio. More KRAS reads were captured from CAsE-PE cDNA compared to RWPE-1 by each probe, and only CAsE-PE cDNA showed the presence of KRAS viral fusion transcripts, primarily mapping to LTR and endogenous retrovirus sequences on either the 5'- or 3'-end of KRAS. A majority of KRAS viral fusion transcripts contained 4 to 6 exons but there were some PacBio sequences with KRAS in unusual orientations, suggesting possible viral insertions within the gene. In other experiments, conditioned media was extracted for retroviral particles and RNA-Seq was performed on the isolates showing KRAS retroviral fusion transcripts in CAsE-PE media only. Truncated KRAS transcripts suggests multiple retroviral integration sites may occur within the KRAS gene producing KRAS retroviral fusions of various lengths. The CAsE-PE in vitro transformation model suggests prime molecular events in arsenic-derived tumors could include an arsenic mutation signature, production of driver oncogenes such as KRAS, activation of endogenous retroviruses with potential for genomic integration, and the formation of retroviral fusion transcripts for gene amplification.